The dilogarithm is a special mathematical function that generalizes the ordinary logarithm and appears throughout number theory, physics, and geometry. Mathematicians have long studied "functional equations" for this function, which are identities showing that a combination of dilogarithm values at different inputs adds up to something simple, often zero or a known constant. One classic building block is Rogers' five-term relation, which links five specific dilogarithm values in a clean algebraic identity. Finding new functional equations, especially ones involving inputs drawn from specific number systems, is genuinely hard and has occupied researchers for decades.
The paper constructs a new functional equation involving five dilogarithm terms whose inputs are expressions built from a single free parameter and a particular square root. The authors derive this identity by combining ten separate applications of the five-term relation, which initially introduces fifty dilogarithm arguments. Through careful algebraic manipulation, 45 of those arguments cancel out, leaving exactly the five that appear in the final identity. The authors also show that this simplification is essentially the only way the construction could have worked, making the result feel natural rather than accidental. The starting point is a pair of integrals that turn out to be equal for a straightforward reason, and the dilogarithm identity is what emerges when you unpack that equality.
Plugging in specific numerical values of the parameter produces a collection of concrete results. Some are identities relating dilogarithm values at numbers drawn from specific quadratic number fields like the rationals extended by the square root of 33 or 17. Others connect values from different number fields or produce new analogues of well-known constants like Catalan's constant. The paper also uncovers "dilogarithm ladders," which are structured sequences of identities tied to algebraic numbers of degree four. Notably, all previously known degree-four ladders involved algebraic numbers whose defining polynomial had only two real roots, while the new ones here have four real roots, making them the first known examples of this totally real type beyond degree three.